Cardio-Ankle Vascular Index Predicts Post-Discharge Stroke in Patients with Heart Failure.

AIM: We aimed to evaluate the significance of the cardio-ankle vascular index (CAVI) to predict stroke in patients with heart failure (HF).
METHODS: This was a prospective observational study, which recruited clinical data from a total of 557 patients who had been hospitalized for HF and undergone CAVI. According to the receiver operating characteristic curve analysis, the accurate cut-off value of CAVI in predicting post-discharge stroke was 9.64. We divided the patients into two groups: the high-CAVI group (HF patients with CAVI ≥ 9.64, n=111, 19.9%) and the low-CAVI group (HF patients with CAVI ＜9.64, n=446, 80.1%). We compared the patients' characteristics and post-discharge prognosis. The primary endpoint was stroke.
RESULTS: The high-CAVI group was older (73.0 vs. 65.5 years old, P＜0.001). Male sex (73.9% vs. 61.4%, P=0.015), coronary artery disease (47.7% vs. 36.1%, P=0.024), and diabetes mellitus (54.1% vs. 37.4%, P=0.001) were more prevalent in the high-CAVI group. In contrast, there was no difference in left ventricular ejection fraction, and prevalence of hypertension and dyslipidemia. The Kaplan-Meier analysis demonstrated that post-discharge stroke rate was higher in the high-CAVI group than in the low-CAVI group (log-rank P=0.005). In multivariate Cox proportional hazard analysis, high CAVI was found to be an independent predictor of stroke, with an adjusted hazard ratio of 3.599, compared to low CAVI.
CONCLUSION: CAVI independently predicts stroke in patients with HF.

The trial registration number: UMIN000029132

Introduction

Atherosclerosis is one of the crucial pathophysiologies of cardiovascular diseases (CVDs), including coronary artery disease, stroke, and heart failure (HF) [1 , 2)] . To date, pulse wave velocity (PWV) has been the gold standard to measure arterial stiffness [3 , 4)] . However, PWV is essentially affected by blood pressure (BP) at the time of measurement [5)] . To overcome this limitation, Shirai et al. have developed a novel index called the cardio-ankle vascular index (CAVI) which, independently of BP, non-invasively represents the stiffness of the aorta, femoral artery, and tibial artery [6)] . The formula of the index is as follows: CAVI=a×[(2ρ/ΔP) ×ln (systolic BP / diastolic BP)×PWV 2 ]＋b, where ρ is blood density, ΔP is pulse pressure, and a and b are coefficients [6)] . CAVI also estimates atherosclerosis in the coronary and carotid arteries more closely than PWV [7 , 8)] .

CAVI is useful not only for the evaluation of arterial stiffness, but also for prognosis prediction in patients who are at high risk of CVDs [9 - 12)] . However, the clinical implication of CAVI in patients with HF has not yet been fully examined, especially regarding prediction of stroke. Thus, the aim of the present study was to evaluate the predictive value of CAVI in terms of stroke in patients with HF.

Methods

Subjects and Protocol

This was a prospective observational study. shows a patient flowchart. Patients were included who (A) were both hospitalized for decompensated HF at Fukushima Medical University Hospital then discharged between March 2010 and September 2019; and (B) underwent CAVI measurement in a stable condition within one week prior to discharge. Decompensated HF was diagnosed on the basis of the current guidelines [2 , 13 , 14)] . A total of 1,242 patients met these criteria. Exclusion criteria included (C) Patients with obvious history of peripheral artery disease and/or atrial fibrillation (including all types: paroxysmal, persistent, long-standing persistent, and permanent atrial fibrillation [ ] ); and (D) those who were receiving maintenance dialysis during the study period. A total of 685 patients were excluded according to these criteria. The definition of peripheral artery disease and atrial fibrillation was in accordance with those used in previous studies [16 - 18)] . Finally, a total of 557 patients were analyzed. The receiver operating characteristic curve analysis revealed that the accurate cut-off value of CAVI in predicting post-discharge stroke was 9.64. We divided patients into two groups based on this cut-off value: the high-CAVI group (patients with CAVI ≥ 9.64, n =111, 19.9%) and the low-CAVI group (those with CAVI ＜9.64, n =446, 80.1%). We compared patient characteristics and post-discharge prognosis between the two groups. The primary endpoint of this study was post-discharge stroke, and we evaluated CAVI as a predictor for this endpoint. Patient characteristics included demographic data at discharge, laboratory and echocardiographic data, and results of CAVI measurement. Laboratory and echocardiographic data were obtained within one week prior to discharge in a stable condition. Estimated glomerular filtration rate (eGFR) was assessed using a three-variable Japanese equation [19)] . The definitions of comorbidities and follow-up methods were in accordance with our previous studies [16 , 17 , 20)] .

Fig.1. Patient flowchart

CAVI, cardio-ankle vascular index.

This study complied with the Declaration of Helsinki and the statement of STROBE (Strengthening the Reporting of Observational studies in Epidemiology) [21 , 22)] . The study protocol was approved by the ethical committee of Fukushima Medical University. All patients gave written informed consent to participate in this study.

Definition of Stroke

Stroke was defined by experienced neurologists in accordance with an established statement as an acute episode of focal dysfunction of the brain, retina, or spinal cord lasting longer than 24 hours, or of any duration if imaging (computed tomography or magnetic resonance imaging) or autopsy showed focal infarction or hemorrhage relevant to the symptoms [23 - 25)] .

CAVI Measurement

CAVI was measured automatically using VaSera VS-1000 (Fukuda Denshi Co., Ltd., Tokyo, Japan) with the patient in the supine position [20 , 26)] . Cuffs were attached bilaterally to the upper arms and ankles. Electrocardiogram electrodes and a microphone were placed on both wrists and on the sternum, respectively. The average values of both sides of CAVI were entered into analyses. The measurement was performed in a stable condition within one week prior to discharge.

Statistical Analysis

All continuous variables analyzed in this study were non-normally distributed according to the Shapiro-Wilk test, and were expressed as medians (25th, 75th percentile). Categorical variables were presented as numbers (percent). Continuous and categorical variables were compared using the Mann-Whitney U test and the chi-square test, respectively. The receiver operating characteristic curve analysis for predicting post-discharge stroke was performed using EZR version 1.40 (Saitama Medical Center, Jichi Medical University, Saitama, Japan) [27)] .We compared the occurrence of post-discharge stroke using the Kaplan-Meier analysis with log-rank test. We assessed CAVI as a predictor for post-discharge stroke using the Cox proportional hazard analysis. To adjust clinical confounding factors, we performed both the subgroup analysis and the multivariate Cox proportional hazard analysis. The univariate Cox proportional hazard analysis was subdivided by subgroups based on presence or absence of categorical factors and the median of continuous variables. Interaction P values were obtained using multivariate model including CAVI, subgroup factors, and interactions between CAVI and subgroup factors. Multivariate Cox proportional hazard analysis was also performed. P values ＜0.05 were considered statistically significant in all analyses. All analyses, except for the receiver operating characteristic curve analysis, were conducted using IBM SPSS Statistics version 26 (IBM, Armonk, NY, USA).

Results

A total of 111 (19.9%) patients belonged to the high-CAVI group. Levels of CAVI were 10.4 (9.9, 11.1) in the high-CAVI group and 7.9 (6.8, 8.7) in the low-CAVI group ( P ＜0.001). Comparisons of patient characteristics between the two groups are shown in . The high-CAVI group was older (73.0 vs. 65.5 years old, P ＜0.001), had a higher prevalence of male sex (73.9% vs. 61.4%, P =0.015), and showed lower levels of body mass index (22.5 vs. 23.8 kg/m 2 , P =0.008) and higher levels of systolic BP (132.0 vs. 124.0 mmHg, P =0.011). In contrast, levels of diastolic BP and the prevalence of New York Heart Association functional class III or IV were equivalent between the two groups. With respect to past medical history, the prevalence of prior stroke was equivalent (18.0% vs. 13.5%, P =0.220), while coronary artery disease (47.7% vs. 36.1%, P =0.024) and diabetes mellitus (54.1% vs. 37.4%, P =0.001) were more prevalent in the high-CAVI group. There were no statistical differences in medication. The high-CAVI group showed higher levels of BNP (235.9 vs. 135.6 pg/mL, P =0.001) and hemoglobin A1c (6.0% vs. 5.7%, P =0.028), and lower levels of hemoglobin (12.5 vs. 13.3 g/dL, P =0.006), eGFR (54.3 vs. 63.4 mL/kg/1.73 m 2 , P ＜0.001), and albumin (3.8 vs. 4.0 g/dL, P ＜0.001). As to echocardiographic findings including left ventricular ejection fraction, stroke volume, and inferior vena cava diameter, there were no statistical differences between the two groups.

Table 1.

Patient characteristics ( n = 557)

	Low-CAVI group ( n = 446)	High-CAVI group ( n = 111)	P value
Demographic data
Age, years old	65.5 (55.0, 75.0)	73.0 (67.0, 80.0)	＜0.001
Male sex, n (%)	274 (61.4)	82 (73.9)	0.015
BMI, kg/m 2	23.8 (21.2, 26.7)	22.5 (20.5, 25.5)	0.008
Systolic BP, mmHg	124.0 (110.0, 141.0)	132.0 (114.5, 149.5)	0.011
Diastolic BP, mmHg	70.0 (60.0, 82.0)	71.0 (61.5, 86.0)	0.186
NYHA functional class 3 or 4, n (%)	16 (3.6)	3 (2.7)	0.456
Etiology of HF			0.055
Ischemic	121 (27.1)	41 (36.9)
Valvular	142 (31.8)	22 (19.8)
Cardiomyopathy	108 (24.2)	30 (27.0)
Others	75 (16.8)	18 (16.2)
Past medical history
Prior stroke, n (%)	60 (13.5)	20 (18.0)	0.220
CAD, n (%)	161 (36.1)	53 (47.7)	0.024
Hypertension, n (%)	313 (70.2)	85 (76.6)	0.182
Diabetes mellitus, n (%)	167 (37.4)	60 (54.1)	0.001
Dyslipidemia, n (%)	336 (75.3)	84 (75.7)	0.941
COPD, n (%)	110 (26.4)	31 (30.4)	0.414
Smoking, n (%)	248 (56.4)	66 (60.6)	0.429
Medication
RAS inhibitors, n (%)	321 (72.0)	82 (73.9)	0.689
Beta blockers, n (%)	324 (72.6)	88 (79.3)	0.154
Loop diuretics, n (%)	248 (55.6)	72 (64.9)	0.077
CCBs, n (%)	162 (36.3)	45 (40.5)	0.411
Anticoagulants, n (%)	198 (44.4)	39 (35.1)	0.077
Antiplatelet agents, n (%)	265 (59.4)	75 (67.6)	0.115
Laboratory data
BNP, pg/mL	135.6 (47.9, 446.7)	235.9 (99.6, 605.3)	0.001
Hemoglobin, g/dL	13.3 (12.1, 14.8)	12.5 (11.3, 13.9)	0.006
eGFR, mL/kg/1.73 m 2	63.4 (50.8, 75.8)	54.3 (40.2, 64.6)	＜0.001
Sodium, mEq/L	140.0 (138.0, 142.0)	140.0 (138.0, 142.0)	0.378
Albumin, g/dL	4.0 (3.7, 4.4)	3.8 (3.3, 4.2)	＜0.001
LDL cholesterol, mg/dL	108.0 (89.0, 129.0)	106.5 (88.0, 136.0)	0.725
HbA1c (JDS), %	5.7 (5.4, 6.3)	6.0 (5.4, 6.7)	0.028
Echocardiographic data
LVEF, %	53.5 (39.5, 64.4)	48.9 (39.6, 57.0)	0.055
Stroke volume, mL	50.0 (38.3, 66.6)	46.7 (36.7, 57.5)	0.112
IVS thickness, mm	10.7 (9.0, 12.3)	10.6 (9.1, 12.0)	0.853
PW thickness, mm	10.6 (9.2, 12.0)	10.7 (9.4, 12.1)	0.611
LAVI, mL/m 2	34.0 (24.1, 49.6)	35.6 (26.2, 51.1)	0.363
RV-FAC, %	41.7 (32.4, 47.8)	42.0 (36.6, 47.7)	0.584
TR-PG, mm	23.0 (17.8, 33.1)	23.1 (18.0, 32.0)	0.630
IVC diameter, mm	13.9 (11.5, 17.0)	13.2 (10.8, 16.6)	0.316

CAVI, cardio-ankle vascular index; BMI, body mass index; BP, blood pressure; NYHA, New York Heart Association; HF, heart failure; CAD, coronary artery disease; COPD; chronic obstructive pulmonary disease; RAS, renin-angiotensin system; CCB, calcium-channel blocker; BNP, B-type natriuretic peptide; eGFR, estimated glomerular filtration rate; LDL, low-density lipoprotein; HbA1c, hemoglobin A1c; JDS, Japan Diabetes Society; LVEF, left ventricular ejection fraction; IVS, interventricular septum; PW, posterior wall; LAVI, left atrial volume index; RV-FAC, right ventricular fractional area change; TR-PG, tricuspid regurgitation pressure gradient; IVC, inferior vena cava.

During the post-discharge follow-up period (median 1415 days), 25 patients reached the primary endpoint (18 ischemic and 7 hemorrhagic stroke). The Kaplan-Meier analysis demonstrated that post-discharge stroke rate was higher in the high-CAVI group than in the low-CAVI group ( , log-rank P =0.005). The unadjusted Cox proportional hazard analysis revealed that high CAVI (vs. low CAVI) was a predictor of post-discharge stroke ( , hazard ratio [HR] 3.015, 95% confidence interval [CI] 1.351–6.727, P =0.007). In addition, there were no interactions between CAVI and all subgroups according to the subgroup analysis ( . Furthermore, because of small event size and to avoid overfitting, we performed multivariate Cox proportional hazard analysis under consideration of confounding factors as much as possible. The predictive value of CAVI was adjusted for three models: age and sex (Model 1); Model 1 plus atherosclerotic risk factors which differed between the groups, namely presence of coronary artery disease and diabetes mellitus (Model 2); and Model 1 plus severity of HF, namely New York Heart Association functional class Ⅲ or Ⅳ, B-type natriuretic peptide, and left ventricular ejection fraction (Model 3). After adjustment for the above confounding factors, high CAVI was an independent predictor of post-discharge stroke ( ; Model 1, HR 2.784, 95% CI 1.168–6.634, P =0.021; Model 2, HR 2.719, 95% CI 1.134–6.518, P =0.025; Model 3, HR 3.599, 95% CI 1.269–10.212, P =0.016).

Fig.2. Kaplan-Meier analysis for occurrence of post-discharge stroke

CAVI, cardio-ankle vascular index.

Table 2.

Cox proportional hazard analysis and the subgroup analysis for predicting stroke (25 events/ n = 557): the impact of high CAVI (vs. low CAVI)

Factor	Subgroup	n	HR	95% CI	P value	Interaction P value
Total	-	557	3.015	1.351–6.727	0.007	-
Age	≥ 68.0	281	2.148	0.737–6.260	0.161	0.327
	＜68.0	276	4.631	1.355–15.828	0.015
Sex	Male	356	2.430	0.863–6.842	0.093	0.381
	Female	201	5.679	1.518–21.253	0.010
BMI	≥ 23.4	273	3.757	1.224–11.534	0.021	0.573
	＜23.4	270	2.076	0.495–8.698	0.318
Systolic BP	≥ 126.0	281	2.769	1.030–7.441	0.044	0.936
	＜126.0	276	2.839	0.709–11.370	0.141
Diastolic BP	≥ 71.0	281	2.710	0.884–8.309	0.081	0.766
	＜71.0	276	3.199	1.015–10.082	0.047
NYHA	1 or 2	538	2.925	1.263–6.775	0.012	0.617
functional class	3 or 4	19	3.266	0.188–56.776	0.417
Prior stroke	Yes	80	2.962	0.770–11.389	0.114	0.781
	No	477	2.927	1.041–8.233	0.042
CAD	Yes	214	1.154	0.306–4.350	0.833	0.066
	No	343	6.318	2.202–18.124	0.001
Hypertension	Yes	398	3.093	1.301–7.357	0.011	0.751
	No	159	1.887	0.196–18.180	0.583
Diabetes	Yes	227	3.428	1.101–10.667	0.033	0.819
mellitus	No	330	2.525	0.777–8.204	0.123
Dyslipidemia	Yes	420	2.044	0.775–5.390	0.148	0.112
	No	137	10.350	1.876–57.108	0.007
COPD	Yes	141	2.735	0.611–12.244	0.188	0.874
	No	378	2.391	0.815–7.015	0.112
Smoking	Yes	314	4.721	1.440–15.476	0.010	0.213
	No	235	1.722	0.463–6.412	0.418
RAS inhibitors	Yes	403	3.689	1.497–9.088	0.005	0.356
	No	154	1.276	0.145–11.258	0.826
Beta blockers	Yes	412	3.876	1.570–9.569	0.003	0.338
	No	145	1.065	0.124–9.115	0.954
Loop diuretics	Yes	320	2.955	1.165–7.492	0.022	0.774
	No	237	2.458	0.473–12.770	0.285
CCBs	Yes	207	2.517	0.957–6.617	0.061	0.808
	No	350	3.248	0.765–13.790	0.110
Anticoagulants	Yes	237	3.976	1.109–14.247	0.034	0.632
	No	320	2.449	0.871–6.887	0.089
Antiplatelet	Yes	340	2.403	0.855–6.756	0.096	0.453
agents	No	217	4.950	1.376–17.812	0.014
BNP	≥ 158.9	245	2.556	0.858–7.613	0.092	0.415
	＜158.9	244	5.089	1.137–22.773	0.033
Hemoglobin	≥ 13.2	261	2.767	0.692–11.070	0.150	0.815
	＜13.2	258	3.700	1.315–10.407	0.013
eGFR	≥ 61.1	258	1.886	0.216–16.512	0.566	0.657
	＜61.1	257	2.844	1.066–7.582	0.037
Sodium	≥ 140.0	312	2.323	0.713–7.565	0.162	0.535
	＜140.0	206	4.279	1.305–14.033	0.016
Albumin	≥ 4.0	259	3.095	0.567–16.908	0.192	0.916
	＜4.0	227	3.543	1.350–9.295	0.010
LDL	≥ 108.0	206	6.303	1.410–28.187	0.016	0.956
cholesterol	＜108.0	205	6.241	1.393–27.956	0.017
HbA1c (JDS)	≥ 5.8	195	11.338	2.283–56.315	0.003	0.377
	＜5.8	194	4.219	0.944–18.867	0.060
LVEF	≥ 52.2	209	1.699	0.360–8.014	0.503	0.209
	＜52.2	209	6.384	2.008–20.302	0.002
Stroke volume	≥ 49.0	213	3.759	0.686–20.597	0.127	0.933
	＜49.0	208	3.082	1.146–8.287	0.026
IVS thickness	≥ 10.6	255	4.342	1.709–11.028	0.002	0.595
	＜10.6	236	2.381	0.435–13.051	0.317
PW thickness	≥ 10.6	251	6.115	2.206–16.952	0.001	0.155
	＜10.6	239	1.358	0.282–6.550	0.703
LAVI	≥ 34.1	180	2.537	0.489–13.171	0.268	0.234
	＜34.1	180	7.746	1.726–34.776	0.008
RV-FAC	≥ 41.8	118	5.079	0.401–64.306	0.210	0.892
	＜41.8	118	6.187	1.213–31.547	0.028
TR-PG	≥ 23.0	188	2.272	0.565–9.134	0.248	0.377
	＜23.0	174	4.486	1.289–15.615	0.018
IVC diameter	≥ 13.9	241	3.244	0.961–10.954	0.058	0.900
	＜13.9	237	4.125	1.383–12.301	0.011

CAVI, cardio-ankle vascular index; HR, hazard ratio; CI, confidence interval; BMI, body mass index; BP, blood pressure; NYHA, New York Heart Association; CAD, coronary artery disease; COPD; chronic obstructive pulmonary disease; RAS, renin-angiotensin system; CCB, calcium-channel blocker; BNP, B-type natriuretic peptide; eGFR, estimated glomerular filtration rate; LDL, low-density lipoprotein; HbA1c, hemoglobin A1c; JDS, Japan Diabetes Society; LVEF, left ventricular ejection fraction; IVS, interventricular septum; PW, posterior wall; LAVI, left atrial volume index; RV-FAC, right ventricular fractional area change; TR-PG, tricuspid regurgitation pressure gradient; IVC, inferior vena cava.

Table 3.

Multivariate Cox proportional hazard analysis for stroke (event n = 25/557)

	HR	95% CI	P value
High CAVI (vs. low CAVI) unadjusted	3.015	1.351–6.727	0.007
High CAVI (vs. low CAVI) Model 1	2.784	1.168–6.634	0.021
High CAVI (vs. low CAVI) Model 2	2.719	1.134–6.518	0.025
High CAVI (vs. low CAVI) Model 3	3.599	1.269–10.212	0.016

HF, hazard ratio; CI, confidence interval; CAVI, cardio-ankle vascular index.

Model 1: adjusted for age and sex.

Model 2: adjusted for age, sex, coronary artery disease, and diabetes mellitus.

Model 3: adjusted for age, sex, New York Heart Association functional class 3 or 4, B-type natriuretic peptide, and left ventricular ejection fraction.

Discussion

To the best of our knowledge, this study was the first to investigate the association between CAVI and post-discharge stroke in hospitalized patients with HF. The main findings of this study were that: (A) patients with high CAVI (≥ 9.64) had several indicators for severity of HF, including higher age, lower body mass index, coronary artery disease, diabetes mellitus, elevated levels of BNP, lower levels of hemoglobin, impaired renal function, and malnutrition; and (B) high CAVI was an independent predictor of stroke in patients with HF.

The main differences between the two groups in this single-center observational study were consistent with the results of the nationwide multicenter registry of patients who were at risk of CVD: CAVI levels were higher in men than in women, and increased according to age [9)] . Atherosclerosis-related diseases were more prevalent in the High-CAVI group. Izuhara et al. reported that CAVI, not PWV, was associated with carotid artery atherosclerosis and multi-vessel coronary artery stenosis in patients with suspected coronary disease [7)] . Although both CAVI and PWV reflect arterial stiffness [6 , 28)] , CAVI may be superior to PWV in patients with CVD including HF in terms of BP independency [29)] because BP dramatically fluctuates in conjunctions with CVD itself and medication for CVD through the clinical course in those population [2 , 13 , 14 , 30)] . Atherosclerosis plays a key role in developing HF [2)] and the high-CAVI group were complicated with severe HF. The lower body mass index in the high-CAVI group suggested not only elevated inflammation and right heart pressure [17 , 31)] , but also muscle decline that is associated with atrial stiffness [32 , 33)] . BNP is a major marker of HF severity [2 , 13 , 14)] , and the authors have recently reported that BNP is a predictor of stroke in patients with HF [25)] . Impaired renal function is one of the important comorbidities of HF [2 , 13 , 14 , 34)] . In addition, Kubozono et al. reported a negative correlation between eGFR and CAVI in the general population [35)] . The main features of patients in the high-CAVI group, such as aging, coronary artery disease, diabetes mellitus, and impaired renal function, are associated not only with HF, but also with stroke [2 , 13 , 14 , 36 - 38)] . Concordant with our results using cut-off value of CAVI of 9.64, a recent review of vascular function has proposed that CAVI ≥ 9.0 as an abnormal high range is a marker of vascular failure [39)] . In addition, it has been reported that diabetic patients with CAVI ≥ 9.0 had more cardiovascular events [40)] , patients with metabolic syndrome and CAVI ≥ 10.0 had a higher incidence of cardiovascular events [41)] , and CAVI ≥ 9.0 was independently associated with rapid decline in eGFR in patients who were at high risk of CVDs [42)] .

The pathological subtypes of stroke are ischemic stroke (cerebral, retinal, and spinal infarction) and hemorrhagic stroke (intracranial hemorrhage and subarachnoid hemorrhage) [24)] . Ischemic and hemorrhagic stroke share a common pathology, but the proportion of pathological and etiological subtypes of stroke vary depending on the populations of different age, race, ethnicity, and country [24 , 38)] . Arteriosclerosis is one of the two main pathological features of cerebral small vessel disease [36)] . Pathological changes in the cerebral small vessels (e.g. loss of smooth muscle cells, lumen restriction, vessel wall thickening, and microaneurysms) lead to both ischemic and hemorrhagic stroke [36)] . Choi et al. recruited the data of individuals who had undergone general health examinations, and found that participants with the highest quartile of CAVI were significantly associated with cerebral small vessel diseases [43)] . Atherosclerosis occurs not only in cerebral small vessels, but also in intracranial and extracranial large vessels, which account for 20% of ischemic stroke cases [44)] . The relationship between carotid artery atherosclerosis and CAVI has been established in various patient populations [7 , 8 , 45 - 47)] . In terms of assessment of intracranial atherosclerotic disease, one currently-used modality is high-resolution magnetic resolution resonance imaging, which can directly visualize the vessel wall permitting evaluation of not only luminal stenosis but also vessel wall pathology [48)] . However, high-resolution magnetic resolution resonance imaging is limited by its cost and availability [44 , 48)] . Considering this limitation, CAVI is less expensive, widely used, and able to be a first step screening. CAVI can also indicate the presence of silent brain infarction [49)] . The present study was the first to find that CAVI was an independent predictor for stroke in patients with HF. From our results, clinicians should check for and control atherosclerotic risk factors in patients with HF, especially in those with high values of CAVI, in order to both predict and prevent stroke.

Study Limitations

The present study has several limitations. First, as a prospective cohort study of a single center with a relatively small number of patients, the present results may not be the representative of the general HF population. Since HF generally have several co-morbidities such as atrial fibrillation or peripheral artery disease, measurement of CAVI in all HF patients may not be necessarily useful for predicting stroke. Second, although we performed both subgroup analysis and multivariate Cox proportional hazard analysis with several confounding factors as much as possible, we could not rule out residual confounding variables, and the differences in the backgrounds between the groups might not be completely adjusted. Third, asymptomatic stroke may have failed to have been detected. Fourth, changes in CAVI through the clinical course were not taken into consideration due to the study protocol. Fifth, the data of PWV and carotid artery ultrasonography were not available in the dataset.

Conclusions

High CAVI is an independent predictor of stroke in patients with HF.

Acknowledgements

The authors thank Ms. Kumiko Watanabe, Ms. Yumi Yoshihisa, and Ms. Tomiko Miura for their technical assistance.

Notice of Grant Support

This study was supported in part by a grant-in-aid for Scientific Research (No. 20K07828) from the Japan Society for the Promotion of Science.

COI

Akiomi Yoshihisa and Tomofumi Misaka belong to the Department of Advanced Cardiac Therapeutics at Fukushima Medical University, which is supported by Fukuda-denshi Co, Ltd. This company is not associated with the contents of this study.